Abstract
Nowadays, the management of digital medical images faces increasing challenges due to the volume, diversity, and need for interoperability between systems. The DICOM standard has become the main format for storing and transmitting medical images, enabling the integration of data from studies such as MRI, CT, and USG. However, its complex structure and the increasing volume of data generated by interconnected devices, including IoT sensors, demand new strategies for efficient storage and retrieval. Clinical databases must support large volumes of heterogeneous data while ensuring fast access, availability, and secure information exchange. This review explores the integration of the DICOM standard with medical database systems, emphasizing the role of sensors as a primary source in clinical data management. The findings aim to support the development of more effective strategies for data retrieval and exchange, such as database fragmentation, to reduce query response times and improve information systems used by healthcare professionals and patients. Additionally, various data sets or benchmarks used in the analyzed studies are described. As a result, two approaches are identified as particularly noteworthy among the reviewed works, serving as a reference for future applications and technological developments in healthcare.
1. Introduction
The digital growth of medicine has placed sensors at the heart of the generation, integration, and exploitation of large volumes of clinical data. In the field of IoT, millions of portable and wearable devices continuously capture physiological variables [1]. Health sensors are smart devices, either wearable or implantable, used for health monitoring. Current advances in wireless communications, coupled with the rapid development of smart biosensors, have enabled continuous and cost-effective monitoring of patient health status for home rehabilitation purposes, as well as the secure transfer of individual health information via wireless networks, directly to a central data repository for clinical evaluation and subsequent analysis [2]. The recovered data enrich electronic health records, promote early detection of adverse events, and lay the foundation for personalized medicine [3]. However, sensors transcend the IoT ecosystem and are fully integrated into high-performance diagnostic equipment that produces highly complex images and signals [4].
Medical data types range from biosignals and medical images to laboratory tests and omics data [2]. Modalities such as digital radiography, CT, MRI, or even flow cytometry leverage increasingly precise sensors, generating information in DICOM, XML, or XSD format with metadata rich in clinical context [5]. The challenge is no longer just capture, but also semantic interoperability and data engineering that allow these heterogeneous sources to be merged into clinical repositories.
This work undertakes an in-depth review of recent literature to highlight the pivotal role sensors play in medical databases, including both IoT environments and traditional diagnostic devices. Particular emphasis is placed on methods that implement strategies to optimize the performance of such databases while adhering to the DICOM standard. Furthermore, the article details the distinctive attributes of each study reviewed, including: (i) whether all components required for full implementation are provided, (ii) the ease with which the proposed features can be deployed, (iii) the specific strategy adopted to enhance database performance, (iv) the technologies utilized during implementation, and (v) the benchmarks employed in the evaluation stage.
Several systematic reviews have addressed related subjects in the domain of healthcare information management. Adewole et al. [6] conducted a comprehensive meta-data analysis of CDRs, examining critical aspects such as architecture, security, AI/ML integration, and regulatory frameworks across African and global contexts. Nevertheless, this work does not address sensor-based data generation or the role of the DICOM standard in clinical database optimization. Edayan et al. [7] performed a systematic review of integration technologies in laboratory information systems, focusing on S2S communication protocols and interoperability frameworks such as HL7 and FHIR, without considering the particular challenges posed by high-volume medical imaging data management. Similarly, Tummers et al. [8] identified features and obstacles of health information systems across multiple stakeholder domains, yet their analysis remains at a high level of abstraction, without addressing specific storage or retrieval strategies for DICOM-formatted data. In contrast, the present work uniquely focuses on the convergence of the DICOM standard, sensor-generated clinical data, and database optimization strategies, particularly fragmentation techniques, offering a more targeted and technically grounded analysis of the current state of medical data management in interconnected healthcare environments.
This article is relevant because it provides an in-depth analysis of cutting-edge work in key areas of medical data storage. Beyond outlining current challenges and advances in this field, it offers a focused assessment of opportunities for developing new methods to improve database performance.
This paper is structured as follows. Section 2 introduces the materials and methods used to conduct the review. Section 3 includes the results regarding the research questions. Section 4 describes the findings, expanding the discussion on the analysis of the works. Finally, Section 5 includes the conclusions of this research and future directions.
2. Materials and Methods
This research is based on the methodological guidelines and recommendations proposed by Page [9] through the PRISMA declaration for the preparation and transparent reporting of systematic reviews, on the conceptual framework of Arksey and O’Malley [10] that defines the essential stages for conducting scoping reviews, and on the operational improvements suggested by Levac, Colquhoun, and O’Brien [11], who refine this framework by detailing iterative processes of searching, extraction, and consultation with interested parties. The authors in [10] present five well-defined steps: (1) defining the research questions, (2) identifying relevant studies, (3) applying explicit inclusion-exclusion criteria to select studies, (4) extracting and charting the data, and (5) collating, summarizing, and reporting the results.
2.1. Research Questions
Seven research questions are presented that support this work, defining the way in which the approaches included in this state-of-the-art review are evaluated, and enrich the discussion regarding the results obtained.
- RQ1. Which databases are most commonly employed to store medical information?
- RQ2. What technologies are used to view or store images in DICOM format?
- RQ3. What strategies are used to improve the performance of medical databases?
- RQ4. What data sets are most commonly used to evaluate approaches dedicated to medical data management?
- RQ5. How is medical information transmitted from sensors included in wearables and imaging machines to a central data repository?
- RQ6. What strategies are used for storing and viewing images in DICOM format in heterogeneous repositories, such as electronic medical records?
- RQ7. What are the main challenges and areas of opportunity in the current medical database landscape for future developments?
2.2. Inclusion and Exclusion Criteria
For the preliminary identification of studies, the specialized repositories that would comprise the search universe were initially defined: IEEE Xplore Digital Library, ScienceDirect (Elsevier), SpringerLink, ACM Digital Library, and MDPI. To expand coverage and detect additional potentially relevant literature, the strategy was complemented with a search in Google Scholar. All searches were conducted using the selected keywords and restricted to the time interval between 2015 and 2025, thus ensuring the timeliness and relevance of the works considered. Although the search was restricted to 2015–2025, to ensure currency, it is acknowledged that foundational database architectures developed before this period, such as early PACS implementations and relational DBMS designs, remain widely deployed in clinical environments. Where relevant, such technologies are discussed in the context of the studies that reference or build upon them, ensuring that historically established approaches are not overlooked in the analysis. The main queries used in all repositories were:
- ‘DICOM’ AND ‘database’ AND ‘sensor’ AND ‘fragmentation’
- ‘DICOM’ AND ‘database’ AND ‘sensor’
It is noted that the keyword ‘fragmentation’ was included in the first query, since, although all the techniques used in the included research will be addressed, the impact of this technique on improving database performance in the analyzed approaches will be observed. Table 1 presents the keywords used and the related concepts that guided the search, not only so that these words were included textually, but also mentioned in the desired context.
Table 1.
Keywords and related concepts.
2.3. Relevant Studies and Application of Selection Criteria
The initial search in the specialized databases retrieved 526 records: 63 from SpringerLink, 46 from ScienceDirect, 69 from the ACM Digital Library, 13 from MDPI, and 335 from IEEE Xplore. To broaden coverage, a complementary search was conducted in Google Scholar, which generated 1090 pages of results (about 10,900 records). For feasibility, only the first 10 pages (102 articles), ordered by publication date, were screened. This screening revealed that all records previously identified in the specialized databases also appeared in Google Scholar, confirming the thoroughness of the initial search. After excluding 14 duplicates, the 102 additional articles from Google Scholar expanded the corpus to 614 unique studies.
The 614 studies were carefully reviewed to discard articles that did not include the related concepts, obtaining 160 studies. A new filter was performed, eliminating studies that did cover the concepts of interest, but whose approaches were related to other topics outside the scope of this research. Only 85 were retained and are included in the scope of this research. Figure 1 describes the stages that filtered the works found in each of the specialized repositories.
Figure 1.
PRISMA diagram flow for obtaining relevant studies [9].
Figure 2 summarizes the annual evolution of the analyzed corpus (2015–2025) using a stacked histogram that breaks down publications by study type. The decision to include journal articles, conference papers, and book chapters reflects the publication landscape of this field: conference proceedings often represent the first formal presentation of innovative technical approaches, while book chapters provide consolidated overviews that complement primary research. Restricting the corpus to journal articles alone would have excluded relevant contributions that, in some cases, preceded or informed subsequent publications. Regarding methodological quality, a general difference was observed: journal articles tend to provide more comprehensive experimental evaluations, while conference papers and book chapters more frequently report prototypes or proof-of-concept implementations with limited validation. This distinction is reflected in the benchmark column of Table 2, Table 3, Table 4, Table 5, Table 6 and Table 7, where the absence of formal evaluation is more prevalent among these latter source types. No language restrictions were applied during the search; however, all retrieved studies were published in English, which may reflect a publication bias rather than an intentional exclusion criterion.
Figure 2.
Type of publication from 2015 to 2025.
Table 2.
Comparison of the aspects of the main studies obtained from IEEE Xplore.
Table 3.
Comparison of the aspects of the main studies obtained from SpringerLink.
Table 4.
Comparison of the aspects of the main studies obtained from ScienceDirect.
Table 5.
Comparison of the aspects of the main studies obtained from ACM Digital Library.
Table 6.
Comparison of the aspects of the main studies obtained from MDPI.
Table 7.
Comparison of the aspects of the main studies obtained from Google Scholar.
At first glance of Figure 2, a gradual increase in the total volume is evident—with notable peaks in 2020 (13 studies) and 2024 (14 studies)—and a shift in the composition of sources: while in 2015–2016 conference papers predominated, from 2017 onward, journal articles clearly became the majority and concentrated the most recent production. Below the x-axis, the figure includes a numerical table that facilitates precise reading of each bar, indicating how many journal articles (green), conference papers (purple), and book chapters (orange) are recorded per year. In aggregate terms, 65.88% of the total corresponds to articles published in journals (56 of 85 studies), 29.41% to conference proceedings (25 studies), and only 4.71% to book chapters (4 studies), which confirms the relevance of the topics considered in this work and the high quality of the papers reviewed in this study.
Figure 3 summarizes the provenance of the 85 primary studies identified in the systematic search. Google Scholar accounts for the largest share of the corpus (37%), confirming its usefulness as a collector of multiple sources, replicating and expanding on the results found in specialized databases. Among the disciplinary repositories, IEEE Xplore contributes 15 works (18%), followed by SpringerLink and ScienceDirect with 12 each (14% respectively), while MDPI adds 9 records (11%), and the ACM Digital Library 5 (6%).
Figure 3.
Number of studies by digital libraries.
Figure 4 summarizes the geographical origin of the studies included in this review. As can be seen, the largest proportion of papers originated in the United States, followed by a European core led by Switzerland, the Netherlands, and the United Kingdom (between 9 and 11 studies each). The contributions of the remaining countries are much more dispersed: only Canada, India, and Ukraine contribute more than two articles, while nations in Asia, Latin America, Africa, and Eastern Europe, for example, Japan, Brazil, and Kenya are represented with one or two studies.
Figure 4.
Geographical distribution of primary studies.
The predominance of studies from the United States and Europe in Figure 4 likely mirrors broader patterns in research output and database infrastructure investment, rather than a limitation of the search strategy itself. This geographical imbalance is acknowledged as a potential constraint on the global generalizability of the conclusions drawn in this review.
2.4. Collating, Summarizing, and Reporting
Once the main articles were obtained for review in this study, different characteristics of each approach were extracted for further analysis. The following sections detail the characteristics evaluated and the findings, continuing the PRISMA process.
3. Results
The studies obtained were reviewed according to six aspects derived from the research questions described above. The aspects are described below:
- Technologies used by each approach to store medical information, to view images in DICOM format, or to manage these same types of images. The technologies identified in each study are reported in the corresponding column. It should be noted, however, that PACS systems and database management systems are not included in this field, as both are addressed separately in dedicated tables within the discussion section, where a more detailed and comprehensive analysis of each is presented.
- The different strategies (if mentioned) to improve the performance of medical databases. The strategies documented in the reviewed studies that seek to optimize the performance of medical databases were analyzed.
- Benchmarks or data sets used to evaluate each approach. The review identified benchmarks or data sets used in the studies to evaluate the effectiveness and performance of the different technological strategies.
- Strategies used to transmit information from sensors to a central data repository. We reviewed the different documented communication strategies for the efficient and secure transmission of data obtained by medical sensors to central clinical information repositories.
- The challenges presented in each study or projects for future research were documented and analyzed, primarily related to technical, operational, and economic aspects.
- Strategies used for storing DICOM images in heterogeneous repositories that include alphanumeric medical information. These approaches stand out for their ability to efficiently manage diverse data in a single environment, addressing interoperability and comprehensive management of medical information in complex clinical systems.
Table 2, Table 3, Table 4, Table 5, Table 6 and Table 7 describe each of the main studies under the aspects already mentioned. The tables organize the studies by the data source from which they were obtained and collectively provide a structured overview of the 86 primary studies included in this review. Across all sources, considerable heterogeneity is observed in the technologies employed, the performance strategies adopted, and the communication mechanisms used to transmit data from sensors to central repositories. While some studies report the use of specific benchmarks or recognized data sets to validate their approaches, a notable subset relies on synthetic workloads, prototype implementations, or qualitative evaluations, reflecting the current lack of consolidated benchmarking standards in the field. Similarly, strategies for handling DICOM images within heterogeneous repositories range from distributed and cloud-based architectures to semantic metadata linkage, underscoring the diversity of solutions currently being explored. The findings extracted from these tables serve as the foundation for the in-depth analysis presented in the following section, where each research question is addressed in detail.
4. Discussion
4.1. RQ1. Which Databases Are Most Commonly Used to Store Medical Information?
The analysis of the reviewed studies reveals a notable diversity in the technologies employed for medical data storage, yet some patterns emerge. Distributed databases and Blockchain-based architectures recur across several contributions. For instance, Pedrosa et al. [17] emphasize the integration of distributed databases with blockchain to enhance data security and traceability in EHR. Similarly, Ismail et al. [15] propose the use of blockchain along with IPFS, reinforcing a trend toward decentralized and immutable storage models.
Another frequently cited component is the adoption of EHR platforms, often augmented with standards such as HL7 or ISO 13606, as seen in the study by Conte et al. [21]. These standards ensure interoperability and semantic consistency across healthcare systems. Complementary technologies like cloud storage, edge computing, and IoT devices are also integrated to support scalable and low-latency access to medical data, as demonstrated by Sondur et al. [19].
Enriching the answer to the research question, Table 8 shows the use of DBMS mentioned or considered in the analyzed works.
Table 8.
DBMS included in the analyzed works.
Table 8 synthesizes the database back ends reported across the corpus and underscores the technological heterogeneity of medical data management. Conventional relational systems (MySQL, PostgreSQL, Oracle, SQL Server, MariaDB, Teradata, SAP HANA) remain foundational for transactional EHR and registry workloads. Document-oriented stores (MongoDB and CouchDB) and wide-column/BigTable-style engines (Cassandra, HBase, Hypertable) are adopted to accommodate semi-structured clinical data and high-throughput ingestion. Overall, the recurrent presence of MongoDB, MySQL, PostgreSQL, and CouchDB, alongside more specialized engines, illustrates how the analyzed works align DBMS choice with diverse data modalities (relational, document, image, and graph) and with priorities of scalability, interoperability, and performance.
Notably, the works of Safaei [33,34] stand out for their specialized focus on PACS. Safaei & HabibiAsl [33] introduce a vertical fragmentation model integrated with indexing engines and query optimizers to reduce retrieval time from PACS databases. Meanwhile, Safaei [34] proposes a hybrid fragmentation scheme combining both vertical and horizontal fragmentation, using feature clustering and metadata indexing to enhance retrieval performance. Both studies underscore the importance of tailoring database strategies to the access patterns and structural properties of medical imaging data, particularly DICOM files. Adding to this picture, Soltanmohammadi et al. [92] apply MD5-hash-based partitioning directly within PostgreSQL, demonstrating that hash-driven fragmentation of structured patient records remains a viable complement to the attribute-based fragmentation strategies discussed above, particularly for transactional RDBMS workloads rather than imaging metadata.
Overall, blockchain, EHR systems, and distributed or fragmented database models emerge as the most consistently utilized technologies. The prominence of PACS-specific approaches in Safaei’s work highlights a complementary trend in optimizing image-based medical data management.
4.2. RQ2. What Technologies Are Used to View or Store Images in DICOM Format?
A wide range of technologies is employed to manage medical images in DICOM format, encompassing both storage and visualization tools. Commonly used storage solutions include PACS, which serve as centralized repositories for DICOM images and often integrate with hospital information systems. Several works have adopted PACS or custom PACS-like systems to archive and retrieve medical images effectively (e.g., [27,32]).
For image viewing, a variety of technologies are used, including dedicated DICOM viewers and web-based tools. Some systems incorporate open-source frameworks such as Orthanc or VTK [14,69], which not only provide storage but also RESTful interfaces compliant with DICOMweb, facilitating image visualization over web applications.
The primary aim of PACS was to create a centralized digital repository for various imaging modalities such as X-ray, CT, MRI, and ultrasound, thereby streamlining radiological workflows and enhancing diagnostic efficiency [94]. Many studies use or consider PACS to manage images in DICOM format (31 studies). However, many approaches create their own PACS implementation, and others do not describe the specific PACS application they use. Table 9 presents the studies that specifically mention the PACS they use or if they developed a new approach. It can be seen that the most widely used PACS is Dicoogle; however, most of the papers that mention the use of PACS do not describe the specific implementation included. Three articles describe their own implementation of PACS functions with different characteristics and environments.
Table 9.
Viewers and PACS systems obtained from the analyzed works.
In [23], highlighted for their contributions, the authors employ a combination of DICOMweb services and cloud-based technologies. These solutions include the use of WADO-RS for image retrieval and visualization directly through web browsers, eliminating the need for specialized desktop applications and supporting mobile access.
Additionally, cloud-based services and hybrid architectures are being explored to enhance scalability and accessibility of DICOM image handling, often integrating with IoT devices and medical sensors to streamline the data flow between acquisition and analysis platforms [60,63,69].
In summary, the technologies used to view or store images in DICOM format range from traditional PACS systems and DICOM viewers to modern, web-enabled platforms and RESTful services like DICOMweb. These solutions reflect the evolution of medical imaging infrastructure towards more flexible, interoperable, and scalable systems.
The prevalence of custom PACS implementations suggests that existing open-source solutions, while functional, do not always meet the specific technical or institutional requirements of individual research contexts, such as integration with proprietary hospital systems, support for non-standard workflows, or deployment in resource-constrained environments. This gap represents an area of opportunity for the open-source community to develop more flexible and modular PACS frameworks that can be adapted without full reimplementation.
4.3. RQ3. What Strategies Are Used to Improve the Performance of Medical Databases?
To enhance the performance of medical databases, researchers have implemented a variety of strategies focused on improving scalability, latency, throughput, and communication efficiency. One notable approach is the use of distributed ledger technologies, such as DAG-based systems, which have been shown to reduce latency and increase throughput in healthcare data exchanges (Saweros and Song [12]). Compression techniques are also employed to optimize storage and transmission, particularly in scenarios involving large imaging data sets such as CT scans (Pole and Shriram [13]).
Pipeline architectures leveraging tools such as Apache NiFi and Kafka enable better scalability and flexibility by streamlining data ingestion and processing across heterogeneous sources (Sreepathy et al. [14]). Furthermore, the adoption of decentralized storage and blockchain frameworks has demonstrated improvements in data integrity and trust when compared to traditional client-server models (Ismail et al. [15]).
Other strategies include the use of lightweight mobile applications and SMS-based systems to simplify communication, especially in low-resource settings where infrastructure may be limited (Latif et al. [16]). These approaches collectively illustrate the diversity of methods used to boost the efficiency and reliability of medical data systems, tailored to various technological and contextual constraints.
A functional comparison of the reviewed approaches reveals meaningful differences across key categories. For DICOM storage and access, Almeida et al. [18] and Galletta et al. [51] demonstrate that distributed NoSQL architectures improve scalability over traditional PACS, though at the cost of increased deployment complexity. For cross-institution data exchange, Pedrosa et al. [17] and Lee et al. [64] adopts blockchain and IPFS to enable secure sharing, while Rinty et al. [38] rely on loosely coupled distributed databases with HL7 standards; the former prioritizes auditability while the latter favors simplicity. For data integrity and access control, blockchain-based approaches such as Ismail et al. [15] and Mohsan et al. [61] outperform traditional models in traceability, but face limitations in throughput when handling large imaging files. Finally, for distributed retrieval performance, Safaei [33,34], Le et al. [63], Galletta et al. [51] and Soltanmohammadi et al. [92] show that fragmentation-based strategies yield measurable query response time reductions compared to non-partitioned baselines, whereas Hadoop-based approaches [45,46] prioritize parallel processing throughput over indexing granularity. Some studies address the adaptation of interoperability strategies in scenarios where medical records are dispersed across multiple formats and locations [12].
To more clearly observe the strategies used to improve performance in each study, different tables (Table 10, Table 11, Table 12, Table 13, Table 14, Table 15, Table 16, Table 17 and Table 18) are presented to classify and group the studies by similar strategies.
Table 10.
Software architectures and patterns.
Table 11.
Data management and storage (cloud/distributed/sharding).
Table 12.
Distributed Computing and Big Data (Hadoop/MapReduce).
Table 13.
Edge/IoT and Efficient Capture.
Table 14.
Standards, Interoperability, and PACS.
Table 15.
Blockchain, DLT, and distributed storage.
Table 16.
AI/Deep Learning and Acceleration of Analysis.
Table 17.
Medical Image/Video Processing and Compression.
Table 18.
Architecture and Scalability.
The strategies identified in Table 10, Table 11, Table 12, Table 13, Table 14, Table 15, Table 16, Table 17 and Table 18 reflect considerable technological diversity in optimizing the performance of medical databases. Data sharding, distributed computing using Hadoop/MapReduce, and the use of microservices and blockchain-based architectures are the most common approaches. Furthermore, the integration of artificial intelligence and deep learning techniques is emerging as a growing trend, complemented by edge computing strategies aimed at reducing the load on central repositories.
Regarding the relevance of traditional fragmentation in the context of cloud-native systems, both approaches are not mutually exclusive. While cloud platforms offer built-in sharding and replication mechanisms, explicit fragmentation strategies, such as those proposed by Safaei et al. [33,34], Galletta et al. [51], Le et al. [63], and Soltanmohammadi [92] remain valuable for optimizing DICOM metadata indexing, query performance, and data dissemination in domain-specific scenarios where generic cloud solutions do not account for the structural properties of medical imaging data.
From a technical standpoint, the works of Safaei & HabibiAsl [33] and Safaei [34] provide the most detailed evidence of fragmentation’s impact on DICOM data management. Vertical fragmentation reduced retrieval time by isolating frequently queried metadata attributes, while hybrid fragmentation further improved throughput by combining attribute-based partitioning with feature clustering. Complementing these efforts, Galletta et al. [51] apply a residue-based fragmentation scheme that splits MRI/DICOM files into redundant chunks distributed across multiple cloud storage providers, prioritizing data availability, reliability, and privacy through obfuscated metadata maps rather than query-time optimization. Meanwhile, Le et al. [63] combine vertical partitioning with a grid-based multi-objective genetic algorithm (NSGA-G) to derive row/column data configurations for hybrid DICOM storage, balancing storage space, monetary cost, and query response time. Although direct quantitative comparisons across studies are limited by the use of different data sets and environments, all four works report measurable reductions in query response time compared to non-fragmented baselines, reinforcing fragmentation as a technically justified strategy for large-scale medical image retrieval. Complementing these attribute-based fragmentation strategies, Soltanmohammadi et al. [92] propose a value-based hash partitioning scheme for structured patient records, combining MD5 hashing for anonymization with modulus-based segment assignment across native PostgreSQL partitions. Unlike Safaei’s vertical/hybrid fragmentation of DICOM metadata attributes or Le et al.’s grid-based multi-objective optimization of hybrid DICOM storage, this approach targets uniform load distribution of identifier-keyed transactional records rather than image-feature clustering. Taken together, these findings reinforce that fragmentation strategy selection should be guided jointly by data type and access pattern: attribute or feature-based fragmentation is better suited to imaging metadata retrieval, whereas hash-based partitioning is better suited to uniform load balancing of high-volume transactional patient records.
Synthesizing the strategies identified across Table 10, Table 11, Table 12, Table 13, Table 14, Table 15, Table 16, Table 17 and Table 18, three broad technological tiers emerge: (i) infrastructure-level strategies, including distributed computing, cloud storage, and fragmentation, which target scalability and query performance; (ii) integration-level strategies, such as microservices, pipeline architectures, and interoperability standards, which address data heterogeneity and system connectivity; and (iii) intelligence-level strategies, encompassing AI, deep learning, and federated learning, which focus on analytical capabilities and autonomous decision support. These tiers are not mutually exclusive; several reviewed works [24,37,38,40,41,46,56,60,80,91] combine approaches from two or more levels, reflecting the increasingly multi-layered nature of medical data management systems.
4.4. RQ4. What Data Sets Are Most Commonly Used to Evaluate Approaches Dedicated to Medical Data Management?
To evaluate proposed approaches in medical data management, researchers frequently rely on simulated data sets, synthetic workloads, real patient records, or prototype implementations in controlled environments. A common practice involves using simulated health records, which allow for the safe testing of data integrity and system behavior without compromising patient confidentiality (Ismail et al. [15]; Pedrosa et al. [17]). Similarly, synthetic data sets provide a means to assess scalability and flexibility under different data loads, as seen in the work of Sreepathy et al. [14] and Sondur et al. [19].
In some cases, real-world data sets are employed. For example, Pole and Shriram [13] used actual CT image sequences to test compression strategies, while Godinho et al. [26] conducted case-based evaluations in a cardiology lab. Other studies such as Ourahmoune et al. [22] built their own data sets from tests with healthy volunteers, reflecting a hybrid approach to data set creation.
A subset of studies relies on prototypes tested with sample data or pilot implementations in clinical settings to validate system feasibility (Saweros and Song [12]; Conte et al. [21]). However, some works notably lack formal benchmarking, relying instead on surveys or qualitative evaluations (Dhayne et al. [20]; Spinsante et al. [25]).
Table 19 is presented, relating the different benchmarks or data sets to the works that use them.
Table 19.
Benchmark or data set.
As can be seen in Table 19, the most used data set to evaluate medical data management systems is TCIA, although it is noted that the vast majority of the works use different data sets. It should be noted that some data sets listed in Table 19, such as stock market or climate datasets, appear unrelated to the medical domain. These were included in the reviewed studies as part of multi-domain ingestion pipeline evaluations, where medical data was processed alongside heterogeneous non-medical sources to demonstrate system scalability and flexibility. Their presence in the table reflects the scope of those specific works rather than a thematic misalignment with this review.
To improve reproducibility, future studies should adopt a minimum set of reporting standards, including: the explicit identification of the data set used (size, modality, and source), the hardware and software environment, and the performance metrics evaluated (e.g., query response time, throughput, storage overhead). Adopting widely recognized data sets such as TCIA as a common baseline would further facilitate meaningful comparisons across approaches.
The inclusion of the MIMIC-IV CXR subset (Dubey and Saxena [93]) further substantiates the prevalence of the MIMIC family of datasets noted in the concluding remarks of this review, and partially addresses the current scarcity of MIMIC-based entries in Table 18.
Beyond reporting standards, the establishment of problem-specific benchmark datasets would significantly advance the field. For instance, a unified dataset for DICOM metadata retrieval performance, another for sensor-to-repository transmission latency, and another for heterogeneous data integration could serve as common baselines across studies. TCIA represents a step in this direction for imaging data, but its scope does not cover the full range of problems addressed in this review, underscoring the need for community-driven efforts to define and maintain standardized evaluation corpora for each research question.
4.5. RQ5. How Is Medical Information Transmitted from Sensors Included in Wearables and Imaging Machines to a Central Data Repository?
The transmission of medical data from wearable sensors and imaging machines to central repositories varies depending on the type of device and system architecture employed. Wearable sensors, which collect physiological data in real time, often transmit information through mobile applications or wireless communication protocols to intermediary devices or directly to cloud-based systems. For example, Ullah et al. [24] explicitly mention the use of wearable sensors for data capture, while Latif et al. [16] describe a mobile-based system that handles basic health data inputs via SMS or mobile apps, facilitating lightweight communication in resource-limited settings.
In more advanced scenarios, ambient and spatial sensors like those used by Spinsante et al. [25] and Ourahmoune et al. [22] capture motion, temperature, or ultrasound data, which are then transmitted via local networks to centralized platforms for processing. These systems are typically part of IoT-enabled architectures that emphasize interoperability and continuous monitoring.
For diagnostic imaging systems such as MRI, CT, or X-Ray machines, transmission often relies on established standards like DICOM, which supports structured file ingestion from PACS. This is evident in the work of Almeida et al. [18], who describe the ingestion of DICOM files from PACS servers, and in Dhayne et al. [20], who focus on standard communication protocols for MRI and X-Ray modalities. Additionally, Godinho et al. [26] illustrate integration of ECG signals and DICOM images in a cardiology lab environment, showing the convergence of multi-modal data transmission.
Some systems opt for broader ingestion pipelines, supporting diverse data sources including sensors, administrative systems, and electronic health records (Sreepathy et al. [14]; Pedrosa et al. [17]), reinforcing the trend toward heterogeneous data integration.
Based on the reviewed studies, three broad categories of sensors were identified: (i) wearable and IoT sensors, which capture physiological variables such as ECG, SpO2, and body temperature and typically transmit data via mobile or wireless protocols [1,12,15,19,24,27,38,40,41,44,50,55,58,60,62,68,72,73,80,88,91]; (ii) environmental and ambient sensors, such as motion or temperature detectors used in assisted living contexts [22,25,53]; and (iii) imaging device sensors, including CT detectors, MRI coils, and ultrasound transducers, which generate high-volume DICOM data transmitted through PACS infrastructures [13,17,18,21,23,26,29,31,32,33,34,36,45,46,47,49,51,59,60,61,63,64,65,69,71,72,75,76,77,78,79,82,86]. This classification highlights that each sensor category interacts differently with DICOM standards and medical databases, and that integration strategies must be tailored accordingly.
Table 20 presents an in-depth analysis of all the works that include or mention how the various devices or sensors communicate and collect data. Some works include more than one form of transmission, or the topic addressed encompasses various sources and forms of communication; in these cases, those described in more detail in each report are included.
Table 20.
Summary of device types, communication media, protocols, and standards used in the selected studies.
In summary, the transmission of medical information from sensors to central repositories involves a combination of mobile and wireless technologies for wearables, and standardized DICOM-based communication for imaging devices. These are often integrated into broader IoT and CPS infrastructures that facilitate real-time data exchange and centralized analysis.
Table 20 shows that DICOM is the predominant standard for transmitting medical image data, adopted across virtually all acquisition modalities analyzed. Wi-Fi and the Internet are the most widely used communication channels, while protocols such as HTTP/REST and HL7 FHIR are prevalent in clinical interoperability environments. Also noteworthy is the growing presence of decentralized technologies such as IPFS and blockchain for secure data transmission, as well as the use of Bluetooth and mobile networks in wearable devices for remote patient monitoring.
4.6. RQ6. What Strategies Are Used for Storing and Viewing Images in DICOM Format in Heterogeneous Repositories, Such as Electronic Medical Records?
In heterogeneous medical repositories, such as EHRs, the integration of DICOM images presents challenges related to format diversity, access control, and data linkage. Various strategies have been proposed to address these challenges and ensure seamless storage and viewing of imaging data. A common approach is the conversion and compression of DICOM images into lighter formats to facilitate integration and visualization. Pole and Shriram [13], for instance, describe a method where CT DICOM scans are converted to YUV format for optimized compression and easier handling within EHRs. Another prominent strategy involves web-based DICOM viewers, allowing clinicians to access images directly from heterogeneous platforms. Nguyen et al. [23] demonstrate the feasibility of viewing DICOM files on smartphones through browser-accessible tools, improving usability and mobility.
Some systems integrate DICOM storage into distributed database structures, such as Cassandra, enabling scalable and secure management. Almeida et al. [18] combine this with access management mechanisms to support diverse medical data formats across a unified repository. Efforts have also focused on linking structured and unstructured medical data, including DICOM, using metadata and semantic processing techniques. Dubey and Saxena [93] illustrate a related pattern at the NoSQL level, embedding CXR image binaries alongside descriptive metadata within a single MongoDB document, simplifying retrieval without requiring joins across separate structured and unstructured stores, though notably storing the images as JPEG rather than preserving the native DICOM container. Conte et al. [21] employ semantic rules to connect structured records with imaging information, while Godinho et al. [26] demonstrate a warehouse model that integrates DICOM with other clinical data streams. In blockchain-based systems, as described by Ismail et al. [15], DICOM images are stored via IPFS, and patient records are securely managed in a decentralized architecture, promoting traceability and access control. Finally, in low-resource contexts, simpler techniques such as text-based encoding of medical information, as seen in the work of Latif et al. [16], are used to maintain some level of interoperability between structured data and image-derived summaries.
Table 9 describes a compendium of viewers and PACS systems found in the studies included in this research.
In summary, strategies for handling DICOM in heterogeneous repositories include format conversion, web-based viewing, distributed and blockchain-backed storage, metadata linkage, and integrated data warehousing. These approaches ensure flexible, scalable, and accessible image management across complex medical data ecosystems.
Table 9 shows that Dicoogle is the most frequently referenced PACS system in the analyzed studies, standing out for offering both storage and visualization capabilities. However, a significant proportion of the studies opt for developing their own implementations, reflecting the need for solutions adapted to specific environments and requirements. Tools such as Orthanc, Cornerstone.js, and XNAT also appear as established alternatives, while several specialized viewers, such as RadiAnt or MicroDicom, are used exclusively for visualization. This diversity underscores the lack of a unified standard in the adoption of PACS systems, representing an area of opportunity for interoperability in heterogeneous clinical environments.
4.7. RQ7. What Are the Main Challenges and Areas of Opportunity in the Current Medical Database Landscape for Future Developments?
Future developments in medical databases are shaped by several recurring challenges and promising research directions. A major area of opportunity lies in enhancing interoperability and infrastructure scalability, which are critical for integrating diverse data sources such as wearables, imaging devices, and hospital systems. Studies highlight the importance of IoT integration and smart contract support to automate and secure data handling (Saweros and Song [12]; Latif et al. [16]).
Another recurring challenge is real-time data ingestion and domain-specific adaptability, particularly in systems dealing with high-volume sensor inputs or heterogeneous medical information. Sreepathy et al. [14] advocate for real-time pipelines and modular data connectors tailored to specific clinical domains. Similarly, Ismail et al. [15] emphasize the need to support a broader spectrum of clinical and administrative data in future solutions.
In terms of performance, there is a recognized need for optimization in indexing, fault tolerance, and hybrid cloud deployments, as highlighted by Almeida et al. [18] and Pedrosa et al. [17]. These improvements aim to balance efficiency with reliability in large-scale or distributed environments.
Several researchers also point to the lack of unified standards and data quality mechanisms, which hinder cross-platform integration and intelligent analytics (Dhayne et al. [20]; Spinsante et al. [25]). Enhancing automation, user customization, and AI support, especially for diagnosis or anomaly detection, has been proposed to address these gaps (Nguyen et al. [23]; Godinho et al. [26]).
Finally, expanding national-level platforms, remote care systems, and patient-centered designs is seen as a frontier for widespread adoption, particularly in contexts aiming for inclusive and scalable health monitoring solutions (Conte et al. [21]; Ullah et al. [24]).
In summary, the future of medical databases is defined by the need for more interoperable, real-time, and intelligent systems. This includes integrating IoT technologies, optimizing architectures for scale and reliability, and aligning infrastructure with patient-centric care and clinical decision support.
The rise of federated learning suggests a gradual shift from purely centralized repositories toward hybrid edge-cloud architectures, where data remains near its source for training purposes while only aggregated model updates are shared centrally. This evolution does not eliminate the need for central repositories, but redefines their role as coordination and governance hubs rather than sole storage points.
The tension between heterogeneous data integration and patient privacy mandates such as GDPR and HIPAA remains a critical challenge. Emerging approaches such as differential privacy, trusted execution environments, and pseudonymisation protocols offer promising paths to reconcile research utility with de-identification requirements, though their adoption in large-scale, multi-institutional deployments is still limited and represents a priority area for future development.
5. Conclusions
Medical database platforms are steadily converging on hybrid, distributed architectures that blend mature relational engines with NoSQL stores, blockchain ledgers, and emerging decentralized-EHR frameworks. This combination provides the elasticity and fault tolerance needed to accommodate highly heterogeneous data (structured clinical notes, real-time physiological waveforms, genomic sequences, and ever-growing DICOM image libraries) while preserving end-to-end auditability. Performance gains are now achieved less through monolithic hardware upgrades and more through in-memory processing, container-oriented microservices, and selective migration to public or hybrid clouds. These strategies compress query latency to near–real time, enabling bedside analytics, continuous ICU monitoring, and AI-assisted decision support. Although DICOM remains the de facto standard for imaging, it is increasingly wrapped in DICOMweb services, RESTful PACS gateways, and object-storage back ends (e.g., S3-compatible buckets), with dynamic transcoding pipelines and HL7-FHIR bridges ensuring multichannel visualization and cross-institution data exchange. On the acquisition side, data from wearables and imaging modalities traverse Bluetooth LE, Wi-Fi 6, 5G, or Zigbee links, funnelled through edge gateways that apply preprocessing, encryption, and rule-based filtering before forwarding to the central repository, offloading core resources, and safeguarding privacy during intermittent connectivity.
Despite these advances, several obstacles persist. Semantic interoperability is still hampered by fragmented ontologies and inconsistent metadata, underscoring the need for unified vocabularies and automated FAIR-compliance workflows. The accelerating adoption of deep-learning pipelines demands infrastructures that can scale horizontally without compromising transactional integrity or regulatory alignment. Privacy and ethics remain front-of-mind, propelling research into federated learning, trusted-execution enclaves, and differential-privacy controls that satisfy both GDPR and HIPAA mandates. Benchmarking, meanwhile, is dominated by the MIMIC and Open-I data sets; broader, openly accessible corpora (particularly in oncology, cardiology, and telemedicine) are essential for rigorous, generalizable evaluations. Looking forward, edge-to-cloud topologies, energy-aware scheduling, and standardized performance metrics (latency, throughput, carbon footprint) will frame the next wave of innovation. Successfully addressing these challenges will pave the way for truly interconnected, patient-centric health ecosystems in which clinical information flows as seamlessly as it is generated, powering safer, faster, and more equitable care.
Author Contributions
Conceptualization, F.C.-M. and L.R.-M.; methodology, F.C.-M. and L.R.-M.; investigation, F.C.-M.; formal analysis, F.C.-M.; data curation, J.A.P.-G.; writing original draft preparation, F.C.-M.; writing review and editing, L.R.-M., G.A.-H., J.A.P.-G., J.C. and J.L.S.-C.; supervision, G.A.-H. and J.C.; project administration, L.R.-M.; funding acquisition, L.R.-M., F.C.-M. and G.A.-H. All authors have read and agreed to the published version of the manuscript.
Funding
This research was funded by Mexico’s Secretariat of Science, Humanities, Technology and Innovation (SECIHTI) through the “Estancias Posdoctorales por México EPM(1) 2024”program (Application No. 8177740), under the project entitled “Nuevo método para la selección y aplicación de técnicas de fragmentación dinámica en bases de datos médicas de gran tamaño”.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
The original contributions presented in this study are included in the article. Further inquiries can be directed to the corresponding author.
Acknowledgments
This work was supported by Mexico’s National Technological Institute (TecNM) and sponsored by both Mexico’s Secretariat of Science, Humanities, Technology and Innovation (SECIHTI).
Conflicts of Interest
The authors declare no potential conflict of interest with respect to the publication of this research.
Abbreviations
The following abbreviations are used in this manuscript:
| AAL | Ambient Assisted Living |
| ACM | Association for Computing Machinery |
| AI | Artificial Intelligence |
| API | Application Programming Interface |
| AWS | Amazon Web Services |
| BCDB | Blockchain Database |
| BFT | Byzantine fault tolerance |
| BIDS | Brain Imaging Data Structure |
| BP | Blood Pressure |
| BSN | Body sensor network |
| CBCT | Cone-Beam Computed Tomography |
| CDA | Clinical Document Architecture |
| C-CDA | Consolidated Clinical Document Architecture |
| CDM | Common Data Model |
| CDR | Clinical Data Repository |
| CNN | Convolutional Neural Networks |
| CPS | Cyber-Physical Systems |
| CPU | Central Processing Unit |
| CR | Cognitive Radio |
| CT | Computed Tomography |
| CXR | Chest X-Ray |
| DAG | Directed Acyclic Graph |
| DB | Database |
| DGT | Discrete Gould Transform |
| DHIS | District Health Information System |
| DICOM | Digital Imaging and Communications in Medicine |
| DLT | Distributed Ledger Technology |
| DR | Digital Radiography |
| DTLZ | Deb–Thiele–Laumanns–Zitzler |
| ECG | Electrocardiogram |
| EDF | European Data Format |
| EEG | Electroencephalogram |
| EHR | Electronic Health Records |
| EMG | Electromyography |
| ERP | Enterprise Resource Planning |
| FAIR | Findable, Accessible, Interoperable, and Reusable |
| FCS | Flow Cytometry Standard |
| FHIR | Fast Healthcare Interoperability Resource |
| FTP | File Transfer Protocol |
| GDPR | General Data Protection Regulation |
| GNU | GNU’s Not Unix |
| GPS | Global Positioning System |
| GSR | Galvanic Skin Response |
| HDFS | Hadoop Distributed File System |
| HDMP | Hospital Data Management Platform |
| HEVC | High Efficiency Video Coding |
| HIPAA | Health Insurance Portability and Accountability Act |
| HIS | Hospital Information System |
| HL7 | Health Level Seven |
| HTTP | Hypertext Transfer Protocol |
| HTTPS | Hypertext Transfer Protocol Secure |
| ICD-10 | International Classification of Diseases, 10th Revision |
| ICT | Information and Communications Technology |
| ICU | Intensive Care Unit |
| IEEE | Institute of Electrical and Electronics Engineers |
| IHE | Integrating the Healthcare Enterprise |
| IMU | Inertial Measurement Unit |
| IoHT | Internet of Healthcare Thing |
| IoMT | Internet of Medical Things |
| IoT | Internet of Things |
| IP | Internet Protocol |
| IPFS | InterPlanetary File System |
| IR | Infrared |
| IRCCS | Istituto di Ricovero e Cura a Carattere Scientifico |
| IS | Information System |
| ISO | International Organization for Standardization |
| IT | Information Technology |
| ITK | Insight Segmentation and Registration Toolkit |
| JADE | Java Agent Development Environment |
| JSON | JavaScript Object Notation |
| JSON-LD | JSON for Linked Data |
| KL | Kellgren and Lawrence |
| LE | Low Energy |
| LSIC | Latin Square Image Ciphe |
| MAM | Masked Authenticated Messaging |
| MDPI | Multidisciplinary Digital Publishing Institute |
| MIDAS | Medical Imaging and Information Datasets |
| MIMIC | Medical Information Mart for Intensive Care |
| ML | Machine Learning |
| MQTT | Message Queuing Telemetry Transport |
| MPI | Message Passing Interface |
| MR | Magnetic Resonance |
| MRI | Magnetic Resonance Imaging |
| NFC | Near Field Communication |
| NLP | Natural Language Processing |
| NoSQL | Not Only Structured Query Language |
| NSE | National Stock Exchange |
| NSGA | Non-dominated Sorting Genetic Algorithm |
| OHIF | Open Health Imaging Foundation |
| OMOP | Observational Medical Outcomes Partnership |
| PACS | Picture Archiving and Communication Systems |
| PBFT | Practical Byzantine Fault Tolerance |
| Portable Document Format | |
| PET | Positron Emission Tomography |
| PHR | Personal Health Record |
| PNT | Personal and Non-Transferable record |
| PRISMA | Preferred Reporting Items for Systematic reviews and Meta-Analyses |
| PSG | Polysomnography |
| PSNR | Peak Signal-to-Noise Ratio |
| PSP | Phosphor Storage Plate |
| PTB | Pulmonary Tuberculosis |
| P2P | Peer-to-Peer |
| QoS | Quality of Service |
| QA | Quality Assurance |
| QAM | Quadrature Amplitude Modulation |
| RDF | Resource Description Framework |
| REST | REpresentational State Transfer |
| RFID | Radio Frequency Identification |
| RIS | Radiology Information System |
| SAREF | Smart Appliances REFerence |
| SHA | Secure Hash Algorithm |
| SMS | Short Message Service |
| SPARQL | SPARQL Protocol and RDF Query Language |
| SSH | Secure Shell |
| SSIM | Structural Similarity Index Measure |
| S2S | Server to Server |
| TCIA | The Cancer Imaging Achieve |
| TCP | Transmission Control Protocol |
| TLS | Transport Layer Security |
| UHPr | Ubiquitous Health Profile |
| UI | User Interface |
| USG | Ultrasound Sonography |
| VTK | Visualization Toolkit |
| WADO-RS | Web Access to DICOM Objects–RESTful Services |
| WfMS | Workflow Management System |
| WPA2 | Wi-Fi Protected Access II |
| WSN | Wireless Sensor Networking |
| XSD | XML Schema Definition |
| XML | Extensible Markup Language |
References
- Moreira, J.; Ferreira Pires, L.; van Sinderen, M.; Daniele, L. SAREF4health: IoT Standard-Based Ontology-Driven Healthcare Systems. Appl. Ontol. 2020, 15, 385–410. [Google Scholar]
- Pezoulas, V.C.; Exarchos, T.P.; Fotiadis, D.I. Types and sources of medical and other related data. In Medical Data Sharing, Harmonization and Analytics; Elsevier: Amsterdam, The Netherlands, 2020; pp. 19–65. [Google Scholar]
- Varma, N.; Han, J.K.; Passman, R.; Rosman, L.A.; Ghanbari, H.; Noseworthy, P.; Avari, J.N.; Deshmukh, A.; Sanders, P.; Hindricks, G.; et al. Promises and Perils of Consumer Mobile Technologies in Cardiovascular Care. J. Am. Coll. Cardiol. 2024, 83, 611–631. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Huang, C.; Wang, J.; Wang, S.; Zhang, Y. Internet of medical things: A systematic review. Neurocomputing 2023, 565, 126719. [Google Scholar]
- Leif, R.C.; Leif, S.H. Cytometry Metadata in XML. In Proceedings of the SPIE BiOS: Imaging, Manipulation, and Analysis of Biomolecules, Cells, and Tissues IX, San Francisco, CA, USA, 13–18 February 2016; pp. 1–7. [Google Scholar]
- Adewole, K.S.; Alozie, E.; Olagunju, H.; Faruk, N.; Aliyu, R.Y.; Imoize, A.L.; Abdulkarim, A.; Imam-Fulani, Y.O.; Garba, S.; Baba, B.A.; et al. A systematic review and meta-data analysis of clinical data repositories in Africa and beyond: Recent development, challenges, and future directions. Discov. Data 2024, 2, 8. [Google Scholar] [CrossRef] [Scilit]
- Edayan, J.M.; Gallemit, A.J.; Sacala, N.E.; Palmer, X.-L.; Potter, L.; Rarugal, J.; Velasco, L.C. Integration technologies in laboratory information systems: A systematic review. Inform. Med. Unlocked 2024, 50, 101566. [Google Scholar] [CrossRef] [Scilit]
- Tummers, J.; Tekinerdogan, B.; Tobi, H.; Catal, C.; Schalk, B. Obstacles and features of health information systems: A systematic literature review. Comput. Biol. Med. 2021, 137, 104785. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Page, M.J.; McKenzie, J.E.; Bossuyt, P.M.; Boutron, I.; Hoffmann, T.C.; Mulrow, C.D.; Shamseer, L.; Tetzlaff, J.M.; Akl, E.A.; Brennan, S.E.; et al. The PRISMA 2020 Statement: An Updated Guideline for Reporting Systematic Reviews. BMJ 2021, 372, n71. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Arksey, H.; O’Malley, L. Scoping Studies: Towards a Methodological Framework. Int. J. Soc. Res. Methodol. 2005, 8, 19–32. [Google Scholar] [CrossRef] [Scilit]
- Levac, D.; Colquhoun, H.; O’Brien, K. Scoping studies: Advancing the methodology. Implement. Sci. 2010, 5, 69. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Saweros, E.; Song, Y.T. Connecting Personal Health Records Together with EHR Using Tangle. In Proceedings of the ACIS International Conference on Software Engineering, Artificial Intelligence, Networking, and Parallel/Distributed Computing (SNPD), Toyama, Japan, 8–11 July 2019. [Google Scholar]
- Pole, A.; Shriram, R. 3D Medical Image Compression By Using HEVC. In Proceedings of the 2018 Fourth International Conference on Computing Communication Control and Automation (ICCUBEA), Pune, India, 16–18 August 2018; pp. 1–5. [Google Scholar]
- Sreepathy, H.V.; Rao, B.D.; Jaysubramanian, M.K.; Rao, B.D. Data Ingestions as a Service (DIaaS): A Unified Interface for Heterogeneous Data Ingestion, Transformation, and Metadata Management for Data Lake. IEEE Access 2024, 12, 156131–156145. [Google Scholar] [CrossRef] [Scilit]
- Ismail, L.; Materwala, H.; Sharaf, Y. BlockHR: A Blockchain-based Healthcare Records Management Framework: Performance Evaluation and Comparison with Client/Server Architecture. In Proceedings of the 2020 International Symposium on Networks, Computers and Communications (ISNCC), Montreal, QC, Canada, 20–22 October 2020; pp. 1–8. [Google Scholar]
- Latif, S.; Rana, R.; Qadir, J.; Ali, A.; Imran, M.A.; Younis, M.S. Mobile Health in the Developing World: Review of Literature and Lessons from a Case Study. IEEE Access 2017, 5, 11540–11556. [Google Scholar] [CrossRef] [Scilit]
- Pedrosa, M.; Lebre, R.; Costa, C. A Performant Protocol for Distributed Health Records Databases. IEEE Access 2021, 9, 125930–125940. [Google Scholar] [CrossRef] [Scilit]
- Almeida, A.; Oliveira, F.; Lebre, R.; Costa, C. NoSQL Distributed Database for DICOM Objects. In Proceedings of the IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Seoul, Republic of Korea, 16–19 December 2020; pp. 1882–1885. [Google Scholar]
- Sondur, S.; Kant, K.; Vucetic, S.; Byers, B. Storage on the Edge: Evaluating Cloud Backed Edge Storage in Cyberphysical Systems. In Proceedings of the IEEE International Conference on Mobile Adhoc and Sensor Systems (MASS), Monterey, CA, USA, 4–7 November 2019; pp. 362–370. [Google Scholar]
- Dhayne, H.; Haque, R.; Kilany, R.; Taher, Y. In Search of Big Medical Data Integration Solutions—A Comprehensive Survey. IEEE Access 2019, 7, 91265–91290. [Google Scholar]
- Conte, R.; Sansone, F.; Grande, A.; Tonacci, A.; Napoli, F.; Pala, A.P.; Raciti, M.; Landi, P. Development of an Integrated ICT System for Data Production, Standardization and Elaboration in Health & Care. In Proceedings of the IEEE E-Health and Bioengineering Conference (EHB), Sinaia, Romania, 22–24 June 2017; pp. 321–324. [Google Scholar]
- Ourahmoune, A.; Larabi, S.; Hamitouche-Djabou, C.; Idri, S.; Benallegue, M. Kinect-based Ultrasound Probe Pose Estimation to Build an Affordable Knee Ultrasound Learning Database. In Proceedings of the International Conference on BioMedical Engineering and Informatics, Shenyang, China, 14–16 October 2015; pp. 489–494. [Google Scholar]
- Nguyen, T.; Nguyen, N.; Pham, B.; Tran, L.; Pham, L.; Le, T. Design of Web based DICOM Processing Software System for Telemedicine with Mobile and Smart Television. In Proceedings of the International Conference on Advanced Computing and Application, Ho Chi Minh City, Vietnam, 27–29 November 2018; pp. 42–49. [Google Scholar]
- Ullah, K.; Shah, M.A.; Zhang, S. Effective Ways to Use Internet of Things in the Field of Medical and Smart Health Care. In Proceedings of the International Conference on Intelligent Systems Engineering (ICISE), Islamabad, Pakistan, 15–17 January 2016; pp. 372–379. [Google Scholar]
- Spinsante, S.; Gambi, E.; Montanini, L.; Raffaeli, L. Data Management in Ambient Assisted Living Platforms Approaching IoT: A Case Study. In Proceedings of the IEEE Globecom Workshops (GC Wkshps), San Diego, CA, USA, 6–10 December 2015; pp. 1–7. [Google Scholar]
- Godinho, T.M.; Almeida, E.; Silva, L.A.; Costa, C. Integrating Multiple Data Sources in a Cardiology Imaging Laboratory. In Proceedings of the International Conference on e-Health Networking, Applications and Services, Munich, Germany, 14–16 September 2016; pp. 1–6. [Google Scholar]
- Syed, L.; Jabeen, S.; Manimala, S.; Elsayed, H.A. Data Science Algorithms and Techniques for Smart Healthcare Using IoT and Big Data Analytics. In Smart Techniques for a Smarter Planet; Mishra, M., Mishra, B., Patel, Y., Misra, R., Eds.; Springer: Cham, Switzerland, 2019; pp. 211–241. [Google Scholar]
- Samson, M.; Swetha, L. Remote Health Care System. In Proceedings of the ICICCT 2019—System Reliability, Quality Control, Safety, Maintenance and Management, Hyderabad, India, 9–11 January 2019. [Google Scholar]
- Praveenkumar, P.; Priya, P.C.; Avila, J.; Thenmozhi, K.; Rayappan, J.B.B.; Amirtharajan, R. Tamper Proofing Identification and Authenticated DICOM Image Transmission Using Wireless Channels and CR Network. Wirel. Pers. Commun. 2017, 97, 5573–5595. [Google Scholar] [CrossRef] [Scilit]
- Ndlovu, K.; Scott, R.E.; Mars, M. Interoperability Opportunities and Challenges in Linking mHealth Applications and eRecord Systems: Botswana as an Exemplar. BMC Med. Inform. Decis. Mak. 2021, 21, 246. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Estrela, L.R.A.; Bueno, M.R.; Azevedo, B.C.; Sousa, V.C.; Guedes, O.A.; Estrela, C. A Novel Methodology for Detecting Separated Endodontic Instruments Using a Combination of Algorithms in Post-Processing CBCT Software. Sci. Rep. 2025, 15, 6088. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Napravnik, M.; Hržić, F.; Tschauner, S.; Štajduhar, I. Building RadiologyNET: An Unsupervised Approach to Annotating a Large-Scale Multimodal Medical Database. BioData Min. 2024, 17, 22. [Google Scholar] [PubMed]
- Safaei, A.A.; HabibiAsl, S. Diamond: Multi-dimensional Indexing Technique for Medical Images Retrieval Using Vertical Fragmentation Approach. J. Supercomput. 2021, 77, 7089–7148. [Google Scholar] [CrossRef] [Scilit]
- Safaei, A.A. Hybrid Fragmentation of Medical Images’ Attributes for Multidimensional Indexing. Clust. Comput. 2022, 25, 215–230. [Google Scholar]
- Pagella Aguero, H. Review of the Current Technologies and Applications of Digital Image Processing. J. Biomed. Sustain. Healthc. App. 2022, 2, 148–158. [Google Scholar] [CrossRef] [Scilit]
- Elloumi, N.; Seddik, H.; Ben Chaabane, S.; Nadra, T. A 3D Processing Technique to Detect Lung Tumor. Int. J. Adv. Comput. Sci. Appl. 2023, 14, 21–35. [Google Scholar] [CrossRef] [Scilit]
- Blazek, P.; Kuca, K.; Krenek, J.; Krejcar, O. Increasing of Data Security and Workflow Optimization in Information and Management System for Laboratory. In Proceedings of the International Work-Conference on Bioinformatics and Biomedical Engineering, Granada, Spain, 20–22 April 2016; pp. 602–613. [Google Scholar]
- Rinty, M.R.; Prodhan, U.K.; Rahman, M.M. A prospective interoperable distributed e-Health system with loose coupling in improving healthcare services for developing countries. Array 2022, 15, 100218. [Google Scholar]
- Akhtar, M.N.; Haleem, A.; Javaid, M. Exploring the advent of Medical 4.0: A bibliometric analysis systematic review and technology adoption insights. Inform. Health 2024, 1, 16–28. [Google Scholar] [CrossRef] [Scilit]
- González Bermúdez, A.; Carramiñana, D.; Bernardos, A.M.; Bergesio, L.; Besada, J.A. A fusion architecture to deliver multipurpose mobile health services. Comput. Biol. Med. 2024, 168, 107756. [Google Scholar]
- Wang, G.; Nurcahyo, A. Designing Personalized Integrated Healthcare Monitoring System through Blockchain and IoT. Procedia Comput. Sci. 2023, 216, 223–232. [Google Scholar] [CrossRef] [Scilit]
- Vesselkov, A.; Hämmäinen, H.; Töyli, J. Technology and value network evolution in telehealth. Technol. Forecast. Soc. Change 2018, 134, 207–222. [Google Scholar] [CrossRef] [Scilit]
- Loomis, P.W.; Reid, J.S.; Tabor, M.P.; Weems, R.A. Dental Identification & Radiographic Pitfalls. In Forensic Odontology; Elsevier: Amsterdam, The Netherlands, 2018; pp. 25–46. [Google Scholar]
- Beier, M.; Jansen, C.; Mayer, G.; Penzel, T.; Rodenbeck, A.; Siewert, R.; Wu, J.; Krefting, D. Multicenter Data Sharing for Collaboration in Sleep Medicine. Future Gener. Comput. Syst. 2017, 67, 466–480. [Google Scholar] [CrossRef] [Scilit]
- Yang, C.T.; Shih, W.C.; Chen, L.T.; Kuo, C.T.; Jiang, F.C.; Leu, F.Y. Accessing Medical Image File with Co-allocation HDFS in Cloud. Future Gener. Comput. Syst. 2015, 43, 61–73. [Google Scholar] [CrossRef] [Scilit]
- Chahal, P.; Pandey, S. An Efficient Hybrid Approach for Brain Tumor Detection in MR Images using Hadoop-MapReduce. In Proceedings of the IEEE/ACM International Conference on Cyber, Physical and Social Computing and Green Computing and Communications (CPSCom/GreenCom), Rhodes, Greece, 2–6 November 2020; pp. 926–931. [Google Scholar]
- Enzmann, D.R.; Arnold, C.W.; Zaragoza, E.; Siegel, E.; Pfeffer, M.A. Radiology’s Information Architecture Could Migrate to One Emulating That of Smartphones. J. Am. Coll. Radiol. 2020, 17, 1299–1306. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Zheng, W. Current Technologies and Applications of Digital Image Processing. J. Biomed. Sustain. Healthc. Appl. 2022, 3, 13–23. [Google Scholar]
- Vázquez-Ingelmo, A.; Sampedro-Gómez, J.; Sánchez-Puente, A.; Vicente-Palacios, V.; Dorado-Díaz, P.I.; Sánchez, P.L.; García-Peñalvo, F.J. A platform for management and visualization of medical data and medical imaging. In Proceedings of the TEEM’20: Eighth International Conference on Technological Ecosystems for Enhancing Multiculturality, Salamanca, Spain, 21–23 October 2020; pp. 518–522. [Google Scholar]
- Song, Y.T.; Pak, J.; Kalabins, A.; Fouché, S. Standard-based Patient-centered Personal Health Record System. In Proceedings of the 11th International Conference on Ubiquitous Information Management and Communication (IMCOM ’17), Beppu, Japan, 5–7 January 2017. [Google Scholar]
- Galletta, A.; Celesti, A.; Tusa, F.; Fazio, M.; Bramanti, P.; Villari, M. Big MRI Data Dissemination and Retrieval in a Multi-Cloud Hospital Storage System. In Proceedings of the 2017 International Conference on Digital Health (DH ’17), London, UK, 2–5 July 2017; pp. 162–166. [Google Scholar]
- Singh, A.K.; Anand, A.; Lv, Z.; Ko, H.; Mohan, A. A Survey on Healthcare Data: A Security Perspective. ACM Trans. Multimed. Comput. Comm. App. 2021, 17, 1–26. [Google Scholar] [CrossRef] [Scilit]
- Gleiss, A.; Lewandowski, S. Removing Barriers for Digital Health Through Organizing Ambidexterity in Hospitals. J. Public Health 2022, 30, 21–35. [Google Scholar]
- Ait Abdelouahid, R.; Debauche, O.; Mahmoudi, S.; Marzak, A. Literature Review: Clinical Data Interoperability Models. Information 2023, 14, 364. [Google Scholar] [CrossRef] [Scilit]
- Cernian, A.; Tiganoaia, B.; Sacala, I.S.; Pavel, A.; Iftemi, A. PatientDataChain: A Blockchain-Based Approach to Integrate Personal Health Records. Sensors 2020, 20, 6538. [Google Scholar] [PubMed]
- Nowakowski, A.Z.; Kaczmarek, M. Artificial Intelligence in IR Thermal Imaging and Sensing for Medical Applications. Sensors 2025, 25, 891. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Koutras, D.; Stergiopoulos, G.; Dasaklis, T.; Kotzanikolaou, P.; Glynos, D.; Douligeris, C. Security in IoMT Communications: A Survey. Sensors 2020, 20, 4828. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Huda, S.; Islam, M.R.; Abawajy, J.; Kottala, V.N.V.; Ahmad, S. A Cyber Risk Assessment Approach to Federated Identity Management Framework-Based Digital Healthcare System. Sensors 2024, 24, 5282. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Althenayan, A.S.; AlSalamah, S.A.; Aly, S.; Nouh, T.; Mahboub, B.; Salameh, L.; Alkubeyyer, M.; Mirza, A. COVID-19 Hierarchical Classification Using a Deep Learning Multi-Modal. Sensors 2024, 24, 2641. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Shakor, M.Y.; Khaleel, M.I. Recent Advances in Big Medical Image Data Analysis Through Deep Learning and Cloud Computing. Electronics 2024, 13, 4860. [Google Scholar] [CrossRef] [Scilit]
- Mohsan, S.A.H.; Razzaq, A.; Ghayyur, S.A.K.; Alkahtani, H.K.; Al-Kahtani, N.; Mostafa, S.M. Decentralized Patient-Centric Report and Medical Image Management System Based on Blockchain Technology and the Inter-Planetary File System. Int. J. Environ. Res. Public Health 2022, 10, 14641. [Google Scholar]
- Khan, M.A.; Khan, J.; Sehito, N.; Mahmood, K.; Ali, H.; Bari, I.; Arif, M.; Ghoniem, R.M. An Adaptive Enhanced Technique for Locked Target Detection and Data Transmission over Internet of Healthcare Things. Electronics 2022, 11, 2726. [Google Scholar] [CrossRef] [Scilit]
- Le, T.; Kantere, V.; D’Orazio, L. Optimizing DICOM data management with NSGA-G. In Proceedings of the International Workshop on Design, Optimization, Languages and Analytical Processing of Big Data (DOLAP), Lisbon, Portugal, 26 March 2019. [Google Scholar]
- Lee, S.; Kim, J.; Kwon, Y.; Kim, T.; Cho, S. Privacy Preservation in Patient Information Exchange (PIE) Systems based on Blockchain: System Design. J. Med. Internet Res. 2022, 24, e29108. [Google Scholar] [PubMed]
- Czelusniak, D.J.; Fuscolim, E.B.; Canciglieri, O., Jr. DICOM Image Management Through Agents Based Systems. In Proceedings of the ICIST 2015—5th International Conference on Information Society and Technology, Belgrade, Serbia, 8–11 March 2015; pp. 468–472. [Google Scholar]
- Udhay, P.; Bhattacharjee, K.; Ananthnarayanan, P.; Sundar, G. Computer-Assisted Navigation in Orbitofacial Surgery. Indian J. Ophthalmol. 2019, 67, 995–1003. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Gyrard, A.; Abedian, S.; Gribbon, P.; Manias, G.; van Nul, R.; Zatloukal, K.; Nicolae, I.E.; Danciu, G.; Nechifor, S.; Marti-Bonmati, L.; et al. Lessons Learned from European Health Data Projects with Cancer Use Cases: Implementation of Health Standards and Internet of Things Semantic Interoperability. J. Med. Internet Res. 2025, 27, e66273. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Schreiweis, B.; Kinast, B.; Ulrich, H.; Pazmino Pinto, S.; Bergh, B. Open Health Knowledge Management Platform: A Comprehensive Evaluation of a Data-centric Approach for Patient Care and Research. Res. Sq. 2024; submitted for publication.
- Vega, F.; Pérez, W.; Tello, A.; Saquicela, V.; Espinoza, M.; Solano-Quinde, L.; Vidal, M.E.; La Cruz, A. WebMedSA: A Web-based Framework for Segmenting and Annotating Medical Images Using Biomedical Ontologies. In Proceedings of the 11th International Symposium on Medical Information Processing and Analysis (SIPAIM), Cuenca, Ecuador, 17–19 November 2015; p. 968110. [Google Scholar]
- Tura, O.; Shabunina, V.; Tur, A. Modern Medical Information Technologies: Implementation Issues and Development Vectors. In Proceedings of the ITTAP 2022: 2nd International Workshop on Information Technologies: Theoretical and Applied Problems, Ternopil, Ukraine, 22–24 November 2022; pp. 1–13. [Google Scholar]
- Pedrosa, M.; Zúquete, A.; Costa, C. A Pseudonymisation Protocol with Implicit and Explicit Consent Routes for Health Records in Federated Ledgers. IEEE J. Biomed. Health Inform. 2021, 6, 2172–2183. [Google Scholar] [CrossRef] [Scilit]
- Alyami, M.A.; Almotairi, M.; Yataco, A.R.; Song, Y.T. Improving Patient Outcomes Through Untethered Patient-Centered Health Records. Adv. Sci. Technol. Eng. Syst. 2018, 3, 164–173. [Google Scholar] [CrossRef] [Scilit]
- Munagandla, V.B.; Pochu, S.; Nersu, S.R.K.; Kathram, S.R. Real-Time Data Integration for Emergency Response in Healthcare Systems. J. AI-Powered Med. Innov. 2024, 3, 25–38. [Google Scholar] [CrossRef] [Scilit]
- Mitrano, G.; Caforio, A.; Calogiuri, T.; Colucci, C.; Mainetti, L.; Paiano, R.; Pascarelli, C. A Cloud Telemedicine Platform Based on Workflow Management System: A Review of an Italian Case Study. Adv. Sci. Technol. Eng. Syst. J. 2022, 7, 87–102. [Google Scholar] [CrossRef] [Scilit]
- Gornale, S.S.; Patravali, P.U.; Hiremath, P.S. A Comprehensive Digital Knee X-Ray Image Dataset for the Assessment of Osteoarthritis. JSM Biomed. Imaging Data Pap. 2020, 6, 1012. [Google Scholar]
- Deniz, Y.; Kaya, S. Determination and Classification of Intraoral Phosphor Storage Plate Artifacts and Errors. Imaging Sci. Dent. 2019, 49, 219–228. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Khoroshun, E.M.; Smelyakov, K.S.; Chupryna, A.S.; Makarov, V.V.; Nehoduiko, V.V.; Vakulik, Y.V. Improving of Computed Tomography Images for Effective Diagnosis of Gunshot Wounds. Emerg. Med. 2024, 20, 577–583. [Google Scholar] [CrossRef] [Scilit]
- Borovska, P.; Ivanova, D.; Draganov, I. Internet of Medical Imaging Things and Analytics in Support of Precision Medicine for Early Diagnostics of Thyroid Cancer. Serdica J. Comput. 2018, 12, 47–64. [Google Scholar] [CrossRef] [Scilit]
- Zimmerer, R.M.; Gellrich, N.-C. Intraoperative Imaging and Postoperative Quality Control. In Facial Trauma Surgery; Dorafshar, A.H., Rodriguez, E.D., Manson, P.N., Eds.; Elsevier: Amsterdam, The Netherlands, 2020; pp. 32–43. [Google Scholar]
- Satti, F.A.; Ali, T.; Hussain, J.; Khan, W.A.; Khattak, A.M.; Lee, S. Ubiquitous Health Profile (UHPr): A Big Data Curation Platform for Supporting Health Data Interoperability. Computing 2020, 102, 2409–2444. [Google Scholar] [CrossRef] [Scilit]
- Podchashynskyi, Y.O.; Stupak, A.G.; Chepiuk, L.O. Analysis of Video Image Compression Methods with Partial Information Loss for Medical Information and Measurement Systems. Tech. Eng. 2024, 2, 199–207. [Google Scholar]
- Molaei, S.; Ghorbani, N.; Dashtiahangar, F.; Peivandi, M.; Pourasad, Y.; Esmaeili, M. FDCNet: Presentation of the Fuzzy CNN and Fractal Feature Extraction for Detection and Classification of Tumors. Comput. Intell. Neurosci. 2022, 2022, 7543429. [Google Scholar] [CrossRef] [Scilit]
- Rasheed, S.; Iqbal, M.S.; Khalid, M.; Sajid, A.; Saeed, M.; Mehmood, A. Revolutionizing Health Care: Upfront Challenges in Implementation of Distributed Healthcare Management Systems. Spectr. Eng. Sci. 2024, 2, 301–332. [Google Scholar]
- Kotzias, K.; Bukhsh, F.A.; Arachchige, J.J.; Daneva, M.; Abhishta, A. Industry 4.0 and Healthcare: Context, Applications, Benefits and Challenges. IET Softw. 2023, 17, 195–248. [Google Scholar]
- Rönnau, R.F.; Rigo, S.J.; Bez, M.; Barbosa, J.L.V. Exploring Heterogeneous Data Processing to Improve Clinical Applications. In Proceedings of the Brazilian Symposium on Computing Applied to Health (SBCAS), Porto Alegre, Brazil, 15–18 September 2020; pp. 108–119. [Google Scholar]
- Bracciale, L.; Loreti, P.; Raso, E.; Bianchi, G. In Plain Sight: A Pragmatic Exploration of the Italian Medical Landscape (In)security. In Proceedings of the Italian Conference on CyberSecurity (ITASEC), Salerno, Italy, 8–11 April 2024; pp. 1–14. [Google Scholar]
- Roehrs, A.; da Costa, C.A.; da Rosa Righi, R.; Silva Farias de Oliveira, K. Personal Health Records: A Systematic Literature Review. J. Med. Internet Res. 2017, 19, e13. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Tiriteu, S.; Cosma, A.; Pacuraru, M.; Zamfir, A.; Chirvase, S. Improved Healthcare Quality Through Integrated Hospital Management and Digitalization. J. E-Health Manag. 2024, 2024, 614161. [Google Scholar]
- Ranschaert, E.R. The Impact of Information Technology on Radiology Services: An Overview. J. Belg. Soc. Radiol. 2016, 100, 93. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Huang, B.E.; Mulyasasmita, W.; Rajagopal, G. The Path from Big Data to Precision Medicine. Expert Rev. Precis. Med. Drug Dev. 2016, 1, 129–143. [Google Scholar] [CrossRef] [Scilit]
- Pendyala, S.K. Cloud-Driven Data Engineering: Multi-Layered Architecture for Semantic Interoperability in Healthcare. J. Bus. Intell. Data Anal. 2023, 1, 1–14. [Google Scholar] [CrossRef] [Scilit]
- Soltanmohammadi, E.; Hikmet, N.; Akgun, D. Tailored Partitioning for Healthcare Big Data: A Novel Technique for Efficient Data Management and Hash Retrieval in RDBMS Relational Architectures. J. Data Anal. Inf. Process. 2025, 13, 46–65. [Google Scholar]
- Dubey, S.A.; Saxena, A. Leveraging MongoDB for Efficient Storage of MIMIC-IV CXR X-ray Images: A Research Perspective. J. Inf. Syst. Eng. Manag. 2025, 10, 613–621. [Google Scholar]
- Pérez-Sanpablo, A.; Quinzaños-Fresnedo, J.; Gutiérrez-Martínez, J.; Lozano-Rodríguez, I.; Roldan-Valadez, E. Transforming Medical Imaging: The Role of Artificial Intelligence Integration in PACS for Enhanced Diagnostic Accuracy and Workflow Efficiency. Curr. Med. Imaging 2025, 21, e18157340. [Google Scholar] [CrossRef] [Scilit]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2026 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license.



